mirror of
https://github.com/apache/httpd.git
synced 2025-08-07 04:02:58 +03:00
Proxy balancer lbmethods are now registered as providers and not
via hooks. Move various find() functions back to mod_proxy_balancer where they belong :) git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@279752 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
@@ -1155,7 +1155,6 @@ PROXY_DECLARE(const char *) ap_proxy_add_balancer(proxy_balancer **balancer,
|
||||
const char *url)
|
||||
{
|
||||
char *c, *q, *uri = apr_pstrdup(p, url);
|
||||
int i;
|
||||
proxy_balancer_method *lbmethod;
|
||||
|
||||
c = strchr(uri, ':');
|
||||
@@ -1173,12 +1172,9 @@ PROXY_DECLARE(const char *) ap_proxy_add_balancer(proxy_balancer **balancer,
|
||||
* NOTE: The default method is byrequests, which we assume
|
||||
* exists!
|
||||
*/
|
||||
lbmethod = (proxy_balancer_method *)conf->lbmethods->elts;
|
||||
for (i = 0; i < conf->lbmethods->nelts; i++) {
|
||||
if (!strcasecmp(lbmethod->name, "byrequests")) {
|
||||
break;
|
||||
}
|
||||
lbmethod++;
|
||||
lbmethod = ap_lookup_provider(PROXY_LBMETHOD, "byrequests", "0");
|
||||
if (!lbmethod) {
|
||||
return "Can't find 'byrequests' lb method";
|
||||
}
|
||||
|
||||
(*balancer)->name = uri;
|
||||
@@ -2069,176 +2065,3 @@ int ap_proxy_lb_workers(void)
|
||||
lb_workers_limit = proxy_lb_workers + PROXY_DYNAMIC_BALANCER_LIMIT;
|
||||
return lb_workers_limit;
|
||||
}
|
||||
|
||||
/*
|
||||
* The idea behind the find_best_byrequests scheduler is the following:
|
||||
*
|
||||
* lbfactor is "how much we expect this worker to work", or "the worker's
|
||||
* normalized work quota".
|
||||
*
|
||||
* lbstatus is "how urgent this worker has to work to fulfill its quota
|
||||
* of work".
|
||||
*
|
||||
* We distribute each worker's work quota to the worker, and then look
|
||||
* which of them needs to work most urgently (biggest lbstatus). This
|
||||
* worker is then selected for work, and its lbstatus reduced by the
|
||||
* total work quota we distributed to all workers. Thus the sum of all
|
||||
* lbstatus does not change.(*)
|
||||
*
|
||||
* If some workers are disabled, the others will
|
||||
* still be scheduled correctly.
|
||||
*
|
||||
* If a balancer is configured as follows:
|
||||
*
|
||||
* worker a b c d
|
||||
* lbfactor 25 25 25 25
|
||||
*
|
||||
* And b gets disabled, the following schedule is produced:
|
||||
*
|
||||
* a c d a c d a c d ...
|
||||
*
|
||||
* Note that the above lbfactor setting is the *exact* same as:
|
||||
*
|
||||
* worker a b c d
|
||||
* lbfactor 1 1 1 1
|
||||
*
|
||||
* Asymmetric configurations work as one would expect. For
|
||||
* example:
|
||||
*
|
||||
* worker a b c d
|
||||
* lbfactor 1 1 1 2
|
||||
*
|
||||
* would have a, b and c all handling about the same
|
||||
* amount of load with d handling twice what a or b
|
||||
* or c handles individually. So we could see:
|
||||
*
|
||||
* b a d c d a c d b d ...
|
||||
*
|
||||
*/
|
||||
|
||||
static proxy_worker *find_best_byrequests(proxy_balancer *balancer,
|
||||
request_rec *r)
|
||||
{
|
||||
int i;
|
||||
int total_factor = 0;
|
||||
proxy_worker *worker = (proxy_worker *)balancer->workers->elts;
|
||||
proxy_worker *mycandidate = NULL;
|
||||
|
||||
|
||||
ap_log_error(APLOG_MARK, APLOG_DEBUG, 0, r->server,
|
||||
"proxy: Entering byrequests for BALANCER (%s)",
|
||||
balancer->name);
|
||||
|
||||
/* First try to see if we have available candidate */
|
||||
for (i = 0; i < balancer->workers->nelts; i++) {
|
||||
/* If the worker is in error state run
|
||||
* retry on that worker. It will be marked as
|
||||
* operational if the retry timeout is elapsed.
|
||||
* The worker might still be unusable, but we try
|
||||
* anyway.
|
||||
*/
|
||||
if (!PROXY_WORKER_IS_USABLE(worker))
|
||||
ap_proxy_retry_worker("BALANCER", worker, r->server);
|
||||
/* Take into calculation only the workers that are
|
||||
* not in error state or not disabled.
|
||||
*/
|
||||
if (PROXY_WORKER_IS_USABLE(worker)) {
|
||||
worker->s->lbstatus += worker->s->lbfactor;
|
||||
total_factor += worker->s->lbfactor;
|
||||
if (!mycandidate || worker->s->lbstatus > mycandidate->s->lbstatus)
|
||||
mycandidate = worker;
|
||||
}
|
||||
worker++;
|
||||
}
|
||||
|
||||
if (mycandidate) {
|
||||
mycandidate->s->lbstatus -= total_factor;
|
||||
mycandidate->s->elected++;
|
||||
}
|
||||
|
||||
return mycandidate;
|
||||
}
|
||||
|
||||
/*
|
||||
* The idea behind the find_best_bytraffic scheduler is the following:
|
||||
*
|
||||
* We know the amount of traffic (bytes in and out) handled by each
|
||||
* worker. We normalize that traffic by each workers' weight. So assuming
|
||||
* a setup as below:
|
||||
*
|
||||
* worker a b c
|
||||
* lbfactor 1 1 3
|
||||
*
|
||||
* the scheduler will allow worker c to handle 3 times the
|
||||
* traffic of a and b. If each request/response results in the
|
||||
* same amount of traffic, then c would be accessed 3 times as
|
||||
* often as a or b. If, for example, a handled a request that
|
||||
* resulted in a large i/o bytecount, then b and c would be
|
||||
* chosen more often, to even things out.
|
||||
*/
|
||||
static proxy_worker *find_best_bytraffic(proxy_balancer *balancer,
|
||||
request_rec *r)
|
||||
{
|
||||
int i;
|
||||
apr_off_t mytraffic = 0;
|
||||
apr_off_t curmin = 0;
|
||||
proxy_worker *worker = (proxy_worker *)balancer->workers->elts;
|
||||
proxy_worker *mycandidate = NULL;
|
||||
|
||||
ap_log_error(APLOG_MARK, APLOG_DEBUG, 0, r->server,
|
||||
"proxy: Entering bytraffic for BALANCER (%s)",
|
||||
balancer->name);
|
||||
|
||||
/* First try to see if we have available candidate */
|
||||
for (i = 0; i < balancer->workers->nelts; i++) {
|
||||
/* If the worker is in error state run
|
||||
* retry on that worker. It will be marked as
|
||||
* operational if the retry timeout is elapsed.
|
||||
* The worker might still be unusable, but we try
|
||||
* anyway.
|
||||
*/
|
||||
if (!PROXY_WORKER_IS_USABLE(worker))
|
||||
ap_proxy_retry_worker("BALANCER", worker, r->server);
|
||||
/* Take into calculation only the workers that are
|
||||
* not in error state or not disabled.
|
||||
*/
|
||||
if (PROXY_WORKER_IS_USABLE(worker)) {
|
||||
mytraffic = (worker->s->transferred/worker->s->lbfactor) +
|
||||
(worker->s->read/worker->s->lbfactor);
|
||||
if (!mycandidate || mytraffic < curmin) {
|
||||
mycandidate = worker;
|
||||
curmin = mytraffic;
|
||||
}
|
||||
}
|
||||
worker++;
|
||||
}
|
||||
|
||||
if (mycandidate) {
|
||||
mycandidate->s->elected++;
|
||||
}
|
||||
|
||||
return mycandidate;
|
||||
}
|
||||
|
||||
/*
|
||||
* How to add additional lbmethods:
|
||||
* 1. Create func which determines "best" candidate worker
|
||||
* (eg: find_best_bytraffic, above)
|
||||
* 2. Create proxy_balancer_method struct which
|
||||
* defines the method and add it to
|
||||
* available server methods using
|
||||
* ap_proxy_add_lbmethods.
|
||||
*/
|
||||
PROXY_DECLARE(int) ap_proxy_add_lbmethods(proxy_server_conf *conf)
|
||||
{
|
||||
proxy_balancer_method *new;
|
||||
|
||||
new = apr_array_push(conf->lbmethods);
|
||||
new->name = "byrequests";
|
||||
new->finder = find_best_byrequests;
|
||||
new = apr_array_push(conf->lbmethods);
|
||||
new->name = "bytraffic";
|
||||
new->finder = find_best_bytraffic;
|
||||
|
||||
return OK;
|
||||
}
|
||||
|
Reference in New Issue
Block a user